Phantosmia: Unveiling Olfactory Hallucinations in Neurology

Introduction: An Alarm Without a Source

Sometime in the early stages of Parkinson’s disease, before the tremors begin, before the shuffling gait appears, before a single clinical diagnosis is made, the brain may start producing smells that have no source. Patients report burning rubber, rotting flesh, cigarette smoke in empty rooms, or sweet chemical odors that appear and vanish without explanation. These experiences are often dismissed as stress, anxiety, or the early signs of psychiatric illness. Families may not be told. Physicians may offer only reassurance. And yet the phenomenon at the center of these experiences, known as phantosmia, is far more medically significant than the curiosity it was long dismissed as.

Phantosmia is the technical term for olfactory hallucination, the perception of a smell in the complete absence of any external odorant. It is not the same as smelling something faint or ambiguous. It is the full, vivid, often intensely unpleasant experience of an odor that exists only within the nervous system. The fact that it can precede the motor symptoms of Parkinson’s disease by years, and that it arises from brain structures now understood to be among the earliest sites of neurodegeneration, has transformed how a small but growing community of neurologists and neuropathologists think about the olfactory system. The nose, it turns out, may be one of the brain’s most sensitive early warning systems. Medicine is only now beginning to listen.

The Neurology Behind Invented Smells

To understand why phantosmia carries such diagnostic weight, it helps to understand what makes the olfactory system anatomically unusual. Unlike every other sensory system in the human body, the olfactory pathway does not route its signals through the thalamus before reaching higher brain centers. Vision, hearing, touch, and taste all pass through this central relay station before being distributed to the cortex. Smell does not. The olfactory bulb, which sits directly above the nasal cavity and projects deep into the temporal lobe, has direct, unmediated access to the limbic system. This architectural shortcut is evolutionarily ancient and shared across most vertebrates. Its consequence is that olfactory signals land immediately in emotional and memory centers, specifically the amygdala and the hippocampus, without first being filtered or contextualized by the brain’s primary sensory relay.

This is why smells trigger memory and emotion with a speed and intensity that other senses rarely match. It is also why, when those same structures malfunction, the results are not merely sensory errors but vivid, emotionally saturated experiences that can feel more real than ordinary perception. Phantosmia does not feel like imagining a smell. It feels like smelling one.

The distinction between phantosmia and parosmia is clinically important and frequently confused. Parosmia involves the distortion of real smells, coffee perceived as sewage being a common example, and is strongly associated with peripheral nerve damage, including the post-viral olfactory injury seen in a significant proportion of long COVID patients. Phantosmia, by contrast, is typically a central nervous system event. There is no external stimulus and no damaged peripheral nerve involved. The hallucination is generated entirely within the brain, most often through spontaneous or dysregulated activity in the piriform cortex, the amygdala, or the uncus of the temporal lobe.

In temporal lobe epilepsy, phantosmia has served as a recognized clinical marker since the nineteenth century. Neurologist John Hughlings Jackson documented olfactory auras in epileptic patients during the 1880s, describing the characteristic burning or sulfurous smells that reliably preceded seizures by seconds to minutes. Modern EEG studies have confirmed that these phantom smells correspond precisely to electrical discharge events in the mesial temporal lobe, particularly in the uncus, a small, hooked structure whose name derives from the Latin for "hook". The association is so consistent that such episodes have come to be called uncinate fits. The olfactory aura, in this context, is not incidental. It is a direct readout of where abnormal electrical activity is occurring in the brain.

Parkinson’s Disease and the Olfactory Early Warning

The relationship between olfactory dysfunction and Parkinson’s disease has become one of the most actively researched areas in neurodegenerative medicine, and the findings have been striking. Studies published in journals including Brain and the Annals of Neurology have found that up to 90 percent of Parkinson’s patients exhibit measurable olfactory loss, a condition called hyposmia, and that this loss often predates the appearance of motor symptoms by four to six years. Phantosmia, the active hallucinatory variant, is less common but has been documented in a meaningful subset of early-stage patients, sometimes emerging as one of the first reported symptoms.

The mechanism centers on alpha-synuclein, the protein that misfolds and aggregates into the Lewy bodies that define Parkinson’s pathology. In 2003, neuropathologist Heiko Braak proposed what became known as the Braak staging hypothesis, a model that fundamentally challenged the prevailing understanding of how the disease progresses. Prior to Braak’s work, Parkinson’s was understood primarily as a disease of the substantia nigra, the midbrain region whose dopaminergic neurons control movement, and whose death produces the motor symptoms most associated with the condition. Braak’s postmortem studies suggested something different: that Lewy body pathology does not begin in the substantia nigra at all. It begins in the olfactory bulb and the dorsal motor nucleus of the vagus nerve, a structure connected to the gut, before spreading upward through the brainstem in a predictable sequence, reaching the motor system only at a relatively advanced stage.

The implications of this model are considerable. If Braak is correct, then olfactory symptoms, including both hyposmia and phantosmia, are not peripheral nuisances or coincidental findings. They are the earliest detectable expression of a disease process that will, years later, disable movement. The substantia nigra, whose damage causes the tremors and rigidity that define clinical Parkinson’s, may be a late casualty of a process that began in the nose. Braak’s model has been substantially supported by postmortem studies and longitudinal imaging research, though some of its finer details remain debated, particularly regarding the role of the gut-brain axis in disease initiation.

Several research groups are now investigating whether standardized olfactory testing could be incorporated into routine screening for populations at elevated genetic risk of Parkinson’s disease, including carriers of mutations in the LRRK2 and GBA genes. The University of Pennsylvania Smell Identification Test, a scratch-and-sniff battery developed in the 1980s, is among the tools being evaluated for this purpose. The appeal is obvious. It is inexpensive, non-invasive, and could potentially identify individuals in a therapeutic window that currently does not exist, a period before motor damage has accumulated to the point of clinical diagnosis.

Brain Tumors, Migraines, and the Psychiatric Overlap

Phantosmia is not exclusive to Parkinson’s disease or temporal lobe epilepsy. It appears across a broader range of neurological and psychiatric conditions, and the diagnostic challenge lies precisely in this breadth. Tumors affecting the frontal and temporal lobes, particularly meningiomas that compress olfactory structures, can present with phantom smells. Olfactory migraines, in which phantom smells serve as the aura preceding a headache, are significantly underdiagnosed and frequently mistaken for anxiety or psychiatric illness. A 2019 review in the journal Cephalalgia estimated that olfactory auras occur in roughly three to four percent of migraine sufferers, a figure likely depressed by underreporting, since patients are often reluctant to describe smelling things that others cannot detect.

The psychiatric overlap is genuinely complex and requires careful clinical navigation. Phantosmia occurs in schizophrenia, where olfactory hallucinations, often of a threatening or persecutory character, are among the most distressing positive symptoms. It has also been reported in severe depression and in the early stages of some dementias, including Lewy body dementia, which shares significant pathological overlap with Parkinson’s disease. The challenge for clinicians lies in distinguishing neurological phantosmia, which typically presents as brief, episodic, and stereotyped, from psychiatric olfactory hallucination, which tends to be more sustained, more emotionally elaborated, and integrated into broader delusional thinking. The smell in a neurological aura is usually the same smell each time. The smell in a psychotic episode is more likely to carry narrative meaning.

Advances in functional MRI and high-resolution olfactory bulb imaging using 7-Tesla scanners are beginning to provide objective correlates that may eventually help resolve this diagnostic ambiguity. Researchers at University College London and at the Max Planck Institute for Empirical Aesthetics have independently published work showing that olfactory bulb volume, measurable by MRI, correlates with both olfactory function and risk of neurodegenerative disease. This suggests that structural imaging of a region long ignored by clinical neuroimaging could become a routine diagnostic tool, offering a window into brain health that current screening protocols largely overlook.

The Emerging Diagnostic Frontier

The broader significance of phantosmia research lies in its implications for the olfactory system as a sentinel for brain health. The nose, for all its evolutionary antiquity and its reputation as the least prestigious of the human senses, turns out to be wired directly into the brain regions most vulnerable to the diseases of aging. Its dysfunction, whether the loss of smell, the distortion of smell, or the invention of smell from nothing, may represent some of the most sensitive early indicators of neurological change that medicine currently possesses.

Several clinical trials are now underway to examine whether olfactory training, which involves repeated and deliberate exposure to strong odorants such as rose, eucalyptus, lemon, and clove, can slow the progression of olfactory loss in early Parkinson’s patients and whether it produces a measurable neuroprotective effect on the olfactory bulb itself. The hypothesis remains speculative, but it rests on a significant biological fact: the olfactory epithelium is one of the few regions of the adult nervous system that continues to generate new neurons throughout life. Stimulating the olfactory pathway may therefore promote neurogenesis in ways that other sensory systems cannot, offering a therapeutic avenue unavailable elsewhere in the brain.

For now, the phantom smells remain medically underappreciated. Most patients who report them are reassured, referred to psychiatry, or simply told there is nothing wrong. The symptoms do not show up on standard blood panels. They do not appear on routine MRI unless someone is specifically looking for olfactory bulb changes. They are easy to dismiss and easy to misattribute. But evidence from neuropathology, functional imaging, and longitudinal cohort studies increasingly suggests that when a patient smells something that is not there, the brain may be trying, in its own distorted and imprecise way, to communicate something that is very real. Learning to interpret that signal, rather than silence it, may prove to be one of the more consequential shifts in early neurological diagnosis of the coming decade.